Two models, two different problems
Most businesses treat their website the way they treat a kitchen renovation. A big capital project every five to seven years, followed by years of pretending nothing is wearing out. That worked when a brochure site was static. It does not work when the thing is your primary sales channel, gets updated by Google every six weeks, and has a dependency tree that drifts whether you touch it or not.
Project budgets and retainers are not alternatives. They are two different financial instruments, and most serious businesses run both. The question is how much of each, and for what.
Project work: capex energy
Project work is bounded scope, fixed outcomes, a defined start and end. Redesign. Replatform. New booking system. Integration with a new CRM. It sits on the balance sheet as a capital-style investment, amortised internally over several years.
It works well when:
- The outcome is defined and measurable.
- The team understands the problem before quoting.
- Nobody is expecting the work to continue after handover.
It works badly when:
- Scope is unclear and the team is under pressure to commit anyway.
- The site is going to need continual attention post-launch (it always does).
- The business cannot distinguish between "build" and "improve" budgets.
The hidden risk of pure project work is feast-famine. You spend AUD 80,000 on a relaunch, nothing for 18 months, then AUD 30,000 in emergency work when a plugin breaks payments. Total spend is higher than a retainer would have been, and the customer experience is worse because nothing compounds between projects.
Retainer work: opex energy
A retainer is a recurring monthly allocation (time, outcomes, or both) applied to the ongoing health and improvement of the site. It sits on the P&L as operating expense. At a minimum it covers keep-the-lights-on work. At its best, it funds a rolling program of performance, conversion and SEO improvements that compound.
It works well when:
- The site is commercially important: lead gen, ecommerce, content.
- The business has a roadmap of improvements, not just tickets.
- The agency brings a point of view about what to prioritise next.
It works badly when:
- It is priced as insurance and delivered as nothing, until something breaks.
- Hours roll over indefinitely or disappear silently.
- There is no reporting, no roadmap, and no meaningful escape clause.
What a good retainer actually delivers
A retainer should pay for itself on the scoreboard, not on the invoice. The categories we budget against, in rough order of impact for most business sites:
Core Web Vitals and performance
LCP, INP and CLS drift every time someone adds a tag, swaps a hero image or installs a plugin. A retainer with a performance line item keeps the site in the green across quarterly Chrome UX Report updates. For ecommerce and content businesses, a 200ms LCP improvement is often worth more than any single feature on the roadmap.
Conversion rate optimisation
Structured A/B testing, funnel analysis, form improvements, copy iteration on top landing pages. Not "redesign the homepage". Small, measured changes to the two or three pages that drive the bulk of revenue.
SEO maintenance
Technical SEO audits, schema updates, content refreshes on ageing cornerstone pages, internal linking hygiene, Search Console triage. The SEO work that does not make the news but keeps a site from quietly sliding down the rankings over 18 months.
Security and dependencies
Framework and plugin updates, PHP and Node version upgrades, SSL and DNS hygiene, periodic dependency audits. This is the category that looks like nothing until the day it looks like everything.
Roadmap delivery
A retainer without a roadmap becomes a helpdesk. A good retainer ringfences hours for planned improvements: the third of the hours that drive the business forward rather than keeping it where it is.
When to pick which
Simple heuristic. If your website directly generates or qualifies leads, sells product, or is a primary content channel, you need a retainer. Size it by the revenue at risk. 5-10% of the annual contribution of the channel is a defensible starting point. Layer projects on top when you have a bounded initiative that would not fit into a monthly cadence.
If your website is a brochure that genuinely does not change, project work with a light maintenance contract is fine. Just be honest with yourself about which category you are actually in. Most "brochure" sites turn out to be lead generation once someone looks at the analytics.
Red flags in retainer pitches
We have reviewed plenty of incumbent retainers when onboarding new clients. The patterns to watch for:
- No monthly reporting. If the agency cannot show you what they did last month and what changed as a result, you are paying for nothing measurable.
- No roadmap. Retainers without a forward plan become reactive. Reactive retainers get cut first in a budget squeeze, and then the site rots.
- Hours that vanish. "Use it or lose it" is fine if the agency proactively surfaces work. It is a scam if the agency waits for tickets and pockets unused time.
- No escape clause. A retainer you cannot exit on 30-60 days notice is a hostage situation. Good agencies do not need lock-in.
- No senior involvement. A retainer staffed entirely by junior resources with no strategic oversight is freelance work at agency rates.
A note on GST
For Australian businesses, standard web retainers from a GST-registered provider include 10% GST and are claimable against your BAS like any other input. Retainers from overseas providers may not include GST, which changes the effective cost comparison and can create reverse-charge obligations with the ATO depending on your setup. Ask your accountant.
Budgeting the mix
A workable starting shape for a serious business website: one major project every 2-3 years, a retainer sized to keep the site competitive in between, and a small discretionary budget for opportunistic work. If you are trying to figure out what that looks like for your business, come and have a chat.
